The knowledge on handover behaviour in WiMAX network is essential for network management and planning in order to achieve optimum system throughput. In this paper we have analysed the handover behaviour of transparent relay in several configurations for the IEEE 802.16j Mobile Multihop Relay (MMR) WiMAX network. The simulation was performed using NCTUns tool and adopted the hard handover mechanism for three different relay network topologies with varying mobile station speeds. The result shows the handover for internal network is faster than the external network and by appropriate relay deployment the system throughput can be increased up to 14.39%.

1. ACEEE Int. J. on Network Security , Vol. 03, No. 02, April 2012
Handover Behaviour of Transparent Relay in WiMAX
Networks
Rohaiza Yusoff1, Mohd Dani Baba2, Muhammad Ibrahim, Ruhani Ab Rahman, Naimah Mat Isa
Centre for Computer Engineering Studies,
Faculty of Electrical Engineering,
Universiti Teknologi MARA, Shah Alam, Selangor, Malaysia.
Email: 1rozzass@yahoo.com, 2mdani074@salam.uitm.edu.my
Abstract-The knowledge on handover behaviour in WiMAX
network is essential for network management and planning
in order to achieve optimum system throughput. In this paper
we have analysed the handover behaviour of transparent relay
in several configurations for the IEEE 802.16j Mobile Multi-
hop Relay (MMR) WiMAX network. The simulation was
performed using NCTUns tool and adopted the hard handover
mechanism for three different relay network topologies with
varying mobile station speeds. The result shows the handover
for internal network is faster than the external network and
by appropriate relay deployment the system throughput can
be increased up to 14.39%.
Figure 1: MMR WiMAX components and applications
Index Terms-WiMAX; Transparent Relay; 802.16j; Hard
Handover II. TRANSPARENT AND NON TRANSPARENT RELAY
I. INTRODUCTION Two different relay modes are defined in the IEEE 802.16j
standard [1] which are the transparent and non-transparent
Handover is a process when a subscriber moves from relays. Transparent relay is usually deployed to enhance the
one connected station to another in the mobility mode. It network capacity while non-transparent relay is used to
was crucial mechanism in order to maintain continuous extend the network coverage. Non-transparent relay does
transmission and avoid any signal distortion. In general, there not forward framing information to BS and this mode has
are two main types of handover which are the vertical or lower complexity and operates for maximum of two-hop
heterogeneous handover and horizontal or homogeneous network topology in the centralized scheduling. The main
handover. Homogeneous handover happens when a purpose of non-transparent relay mode is to extend the BS
subscriber moves within the same network architecture such coverage range by generating Relay Station (RS) own framing
as in the same WiMAX (Worldwide Interoperability for information to forward the package information from BS to
Microwave Access) network. While vertical handover will MS. The throughput enhancement is limited and it can operate
occurs when the subscriber roams from one different network in more than two-hop relay topology with higher complexity.
access to another such as from WiMAX to Wi-Fi networks. Two different procedures in resource allocation and
There are various handover mechanisms available to achieve scheduling are specified for each relay type. Transparent
the maximum quality of service (QoS) in telecommunication relay mode will be scheduled by BS (centralized) and non-
industry. The detail is described in the methodology section. transparent relay mode is scheduled either by BS or RS
IEEE 802.16j standard is for mobile multi-hop relay (distributed) [3]. The success of a good mobility framework
specification for WiMAX. As the mobile technology largely depends on the capability of performing fast and
advances, multi-hop relay specification for WiMAX is seamless handovers irrespective of the deployed architectural
introduced to overcome the demand for coverage extension scenario [4]. The next section will discuss some recent
and capacity enhancement without the need to deploy proposal on handover from various factors.
additional fix backhaul infrastructure. The use of relay in
shadowed area or in-building will provide good coverage III. PREVIOUS WORK
and can maximize the throughput. The many advantages of
the MMR WiMAX approach are less complex in design and This section discusses few recent works on handover
construction than the typical base station, lower total system approaches. For vertical handover between WiMAX and
cost, and enabling the rapid deployment. Multiple hops WLAN, authors [5] have proposed the scheme which
between a Mobile Station (MS) and Base Station (BS) are allocates Mobile Agent Identifier as the authentication,
supported [1]. The components and applications of MMR authorization and accounting (AAA) key during the initial
WiMAX system are as shown in Figure 1 [2]. connection establishment in WiMAX or WLAN. The same
© 2012 ACEEE 11
DOI: 01.IJNS.03.02. 39

2. ACEEE Int. J. on Network Security , Vol. 03, No. 02, April 2012
architecture was also explored by [6] using hybrid technique comprising of two BSs and three RSs with MS7 roaming with
where both the operator and user satisfaction are considered. various speeds travelling from RS5 cell to RS6 cell coverage.
The result from both approaches show lower handover Initially MS7 is connected to BS3 through RS5 and will moves
latency is achieved as compared to the existing technique. from RS5 towards RS6. Three different relay network
To evaluate the proposed algorithm, which is based on multi- configurations are considered for three scenarios A, B and C.
criteria evaluation and analysis can be performed using the
network simulator [6].
Self-Optimization Handover Mechanism [7] which uses
Global Positioning System (GPS) can navigate the system
and combine with the mobility characteristics of mobile RS
as the basic concept in Self-Optimization of Organic
Computing of the mechanism. By using this mechanism, a BS
can be planned in advance for the path choice. To reduce the
number of handover, the channel is scanned to provide a
more stable WiMAX services. Authors [8] designed an RS
grouping algorithm to minimize handoffs by utilizing greedy Figure 2: Topology drawn to observe the handover between two
grouping policy; RS pairs with higher handoff rates will have networks
higher priority for selection. The simulation result shows that In order to match the OFDMA standard [1],[3] the WiMAX
the handoff frequency of the considered MR network can be system model was being set up according to OFDMA
significantly be reduced, and the suitable RS grouping parameters such in Table I.
patterns can be derived using their grouping algorithm.
TABLE I: OFDMA PARAMETER USED
While [9] introduced a new fast handover scheme that able
to maintain a stable data transmission and outperform the
conventional method.
A few analysis was carried out as in [10] and [11] for the
behaviour of handover in Mobile WiMAX environment. The
critical parameters considered are values of the contention
window minimal and maximal size, number of transmission
allocations per frame, number of available CDMA ranging
codes, timers and maximum retransmission limits used in the
processes of ranging, basic capabilities negotiation and
registration. In addition [11] introduced the reduction of
handover delay for certain noise levels by introducing a
A. Types of Handover
continuous scanning process. Another evaluation from [12]
using association and handover optimizations through BS- There are three types of handover mechanism as stated
to-BS signalling the network re-entry time can be significantly in IEEE 802.16j standard, namely the hard handover, macro
shortened. However, the throughput analysis shows that diversity handover (MDHO) and fast base station switching
association has only small effect on throughput; the Mobile (FBSS) [14],[15]. Since the hard handover is mandatory while
IP tunnel establishment was found to be the major MDHO and FBSS are optional, the current public domain
contributor. NCTUns tool only supports hard handover mechanism. The
The National Chiao Tung University – network simulator hard handover mechanism employs the principle of break-
(NCTUns) module has implemented hard handover using the before-make. MS will break the connection with the original
corresponding weight of each link connection. In our study, BS before making a new connection with another BS.
the behaviour of hard handover of transparent relays in Although it may lower the handover quality, the MS need
WiMAX network for the RS – BS – BS – RS configurations not maintain a list of BSs and waste the radio resources.
will be analysed. Figure 3 shows the hard handover mechanism.
IV. METHODOLOGY
In this research, we adopt the free based simulator that
supports relay architecture (IEEE 802.16j) from NCTUns tool.
It provides real-life simulation with re-entering kernel method
and come with friendly user graphical interface (GUI) suites
for beginners [13].
The WiMAX topology was constructed as shown in
Figure 2 for transparent relay types to investigate the effect
Figure 3: Mechanism of hard handover [15]
of handover on throughput for various configurations
© 2012 ACEEE 12
DOI: 01.IJNS.03.02. 39

3. ACEEE Int. J. on Network Security , Vol. 03, No. 02, April 2012
B. MS Path Selection Algorithm
To decide and assign the access station for a MS,
NCTUns module implements a path selection algorithm by
assigning weight for each link connection. The BS decides
the MS’s access station according to the weight of the link
between the MS and RS and the weight of the link between
the RS and BS. The weight of the link corresponds to the
modulation and coding rate used on that link for the required
coverage as shown in Figure 4 which proportional to distance
from the BS. The BS will choose to use the RS as the MS’s
access station if the following condition is satisfied: Figure 5: Graph of handover based on weighted value
Wr + Wp < Ws
process of path link selection or handover is based on the
where,
corresponding weight of each connection. The handover time
Ws: the corresponding weight of MS to BS’s uplink
to transfer from BS3 to BS4 is about 0.3388 ms for all speed as
Wr: the corresponding weight of MS to RS’s uplink
shown in Figure 6 below. The different is very small. This
Wp: the corresponding weight of RS to BS’s uplink
result is also supported by [16] and [17] which have small
changes while the speed or velocity increases.
Figure 4: Relationship between Modulation, Coding and
Coverage.[11]
The speed of MS is also varied and MS7 is set to 5 m/s, 10
m/s, 15 m/s, 20 m/s, 25 m/s and 30 m/s as to observe the effect Figure 6: Graph of handover time based on speeds
of different mobility speeds on handover latency and on
Simulation is also performed to study the effect of handover
network throughput. on throughput for the downlink transmission. The handover
behaviour for BS to RS and from RS to another BS connection
V. RESULT AND DISCUSSION is observed in order to compare their performance.
The results obtained through NCTUns simulations by The system model employs the 16QPSK ¾ modulation
recording the corresponding weight of each MS position per with 18 bytes slot size for 620 available slots in downlink
second. Each connection weight is compared and the lower transmission with a frame duration of 5ms [15]. The theoretical
weight is selected by BS as the communication link. The throughput can be calculated by using the following equation
bandwidth is set to 50 Mbps with MS7 moved at 10 m/s [18].
speed or 36 km/j. Host1 transmits the data from 3 to 120
seconds to MS7 and the path link selection is determined by
the BS. Figure 5 shows the recorded weight for RS5, RS8 and
BS3 for the first 80 seconds. At starting point, MS7 is
connected through RS5 since it has the lowest weight
compared to BS3 and RS8. Furthermore the location of RS5 is From (1) the maximum throughput computed is 17.86 Mbps
the nearest to MS7. When MS7 moves from RS5 towards for downlink transmission over the IEEE 802.16j environment.
RS6, the recorded weight of RS5 will be increasing. When the Host1 sends data continuously to MS7 from 3rd to 120th
weight of direct access BS3 is less than RS5, then the seconds. Figure 7 shows the graph of throughput received
connection will handoff to link BS3. From the graph, the by MS7 while moving from originally connected RS5 to BS3
handoff of MS7 to BS3 by RS5 occurs at the 29th seconds. then to RS8 and finally handover to RS6.
MS7 is served by BS3 about four seconds before Three different RS placements or topologies were de-
handover the connection to RS8. Starting from 33rd seconds ployed to observe the effect on handover. The first topology
to 77th seconds MS7 is being connected through RS8. It is is as shown in Figure 2 where RS8 is the intermediate RS
because MS7 is moving towards RS8 and the weight between BS3 and BS4 and was set up as subnet of BS3. This
decreases. At the 77th seconds, MS7 is handoff to BS3 since topology is referred as Scenario A. Another topology is Sce-
the coverage of BS3 is higher than RS8 before passing MS7 nario C with RS8 being the subnet of BS4. The last Scenario
to neighbouring network, BS4 through RS6. Overall, the B does not have RS8 between BS3 and BS4 as the controlled
© 2012 ACEEE 13
DOI: 01.IJNS.03.02.39

4. ACEEE Int. J. on Network Security , Vol. 03, No. 02, April 2012
simulation. All the throughputs were plotted in the same graph
to facilitate for comparison purposes.
Figure 9: Graph of throughput for different speed of MS7
The simulation of the handover process for transparent relay
mode indicates that for within the same network or cell, the
Figure 7: Graph of throughput for MS7 for scenario A, B and C
handover occurs smoothly without affecting much of the
throughput. However the network throughput decreases
As the first handover occurs on 29 th second when the during the hard handover for different networks due to losses
transmission of MS7 changes from RS5 to BS3, it causes of few packets. Even though the time of handover process is
higher throughput transition and no data lost during this about 0.3388 ms, the change of path link causes the break in
handoff. There is no different in throughput performance for transmission during that short period. The deployment of
the initial stage for all scenarios but after a few seconds, intermediate RS as in Scenario A is able to increase the system
Scenario A throughput begins to drop as it was being handoff throughput up to 14.39%. As stated in [19] the handover
to RS8. While Scenario B and C still maintain at high failure will increase when the velocity of MS increases, thus
throughput until it being handover to BS3 at the 50th seconds. the implementation of RS will helps the situation to minimize
For this second handover, the throughput drops rapidly since the failure.
MS7 is moving away from BS3. Another factor which affects the signal loss is the system
For Scenario A, MS7 has RS8 to enhance its throughput path loss which also influences the process of handover. In
and is able to maintain it until the 77th seconds. After that for this simulation study, the Cost Hatta propagation path loss
all scenarios MS7 is being served by BS3 until it being model is used to simulate the WiMAX propagation for large
transferred to BS4 network. During the process of hard
coverage area.
handover between two networks, the throughput drops until
zero and no data received by MS7. With the aid of RS8,
Scenario A has fewer periods of zero throughputs and after VI. CONCLUSION & FUTURE WORK
that all scenarios show the throughput to increase as usual This paper has presented the throughput of MMR
while connected to BS4 network. WiMAX network which can be enhanced through the
The observation on the effect of different mobile speed appropriate relays deployment. The simulation result shows
on throughput is as in Figure 8 and Figure 9 when using that transparent relay is capable to increase the throughput
topology as in Scenario A. With different speed, the MS7 by 14.39%. The handover time in the same network or cell
handover occurs at different time. The faster the speed the coverage does not affect the throughput however for inter-
earlier the handover occurs. Two graphs are drawn for cell handover the vendor has to plan well in advance for the
convenient to evaluate them since it have different speed appropriate deployment of the MMR WiMAX BS and RS’s
range. The simulation run for 5 m/s mobile speed is extended for optimum system performance. The suitable placement of
to the 200 seconds simulation since the handover occurs at intermediate RS is necessary to minimise the effect of packet
the 174th seconds. The slower the mobility speed, the observed loss or latency due to handover. With higher MS mobility
throughput is more stable. speed, the probability of handover to occur is decreased and
thus proper planning and network implementation is needed
to avoid data losses or an enhancement of system is required.
The selection of relay types and modes is also crucial in
the management of fairly connected network for maximum
system throughput. Thus for future research work it can be
expanded to consider the effect of scheduling of several MS,
directional antenna, path selection or frequency reuse in
specific topology on the overall system throughput.
VII.ACKNOWLEDGEMENT
The authors would like to thanks MOHE, Malaysia and
Figure 8: Graph of throughput for different speed of MS7 UiTM for supporting the study by providing the research
© 2012 ACEEE 14
DOI: 01.IJNS.03.02. 39

5. ACEEE Int. J. on Network Security , Vol. 03, No. 02, April 2012
grants 600-RMI/ST/FRGS 5/3/Fst and 600-RMI/ST/DANA 5/ [10] P. Latkoski and B. Popovski, “Evaluation of Mobile WiMAX
3/Dst respectively handover procedure,” IEEE Mediterranean Electrotechnical
Conference, MELECON2010, pp. 476-481.
[11] K. Daniel, S. Rohde, S. Subik, and C. Wietfeld,
REFERENCES
“Performance Evaluation for Mobile WiMAX Handover with a
[1] IEEE, “Part 16: Air interface for fixed and mobile broadband Continuous Scanning Algorithm,” IEEE Mobile WiMAX Symposium,
wireless access systems: Multi-hop relay specification,” IEEE 2009. MWS ’09. pp. 30-35.
Baseline Document for Draft Standard for Local and Metropolitan [12] M. Majanen, P. H. J. Perala, T. Casey, J. Nurmi, and N.
Area Networks, ed, 2007. Veselinovic, “Mobile WiMAX Handover Performance Evaluation,”
[2] B. Kwon, Y. Chang, and J. A. Copeland, “A Network Entry Fifth International Conference on Networking and Services ICNS
Protocol and an OFDMA Symbol Allocation Scheme for Non- 2009, pp. 263-269.
Transparent Relay Stations in IEEE 802.16j MMR Networks,” [13] S.-Y. Wang, C.-L. Chou, and C.-C. Lin. The GUI User Manual
Military Communications Conference, 2008. for the NCTUns 6.0 Network Simulator and Emulator, 2010.
[3] S. W. Peters and J. Robert W. Heath. The Future of WiMAX: [14] A. G. Davalos, L. Escobar, A. Navarro, and A. Arteaga, “Vertical
Multihop Relaying with IEEE 802.16j. IEEE Communications Handoff Algorithms – A New Approach for Performance
Magazine. 104-111, 2009. Evaluation,” IEEE Globecom 2010 Workshop on Ubiquitous
[4] S. K. Ray, K. Pawlikowski, and H. Sirisena, “Handover in Computing and Networks, 2010.
Mobile WiMAX Networks: The State of Art and Research Issues,” [15] S.-Y. Wang, H.-Y. Chen, and S.-W. Chuang. NCTUns Tool for
Communications Surveys & Tutorials, IEEE, vol. 12, pp. 376-399, IEEE 802.16j Mobile WiMAX Relay Network Simulations, 2010.
2010. [16] S. Murawwat and T. Javaid, “Speed & service based handover
[5] M. J. Yu and S. G. Choi, “A New Mechanism for Fast Handover mechanism for cellular WiMAX,” International Conference on
between Heterogeneous Networks,” ICACT 2011, 2011. Computer Engineering and Technology ,ICCET 2010, pp. V1-418-
[6] M. Shakil, N. Zorba, C. Adam, and C. Verikoukis, “WLAN- V1-422.
WiMAX Vertical Handover Hybrid Satisfaction Mechanism,” IEEE [17] V. Rakovic, V. Atanasovski, and L. Gavrilovska, “Velocity
International Workshop on Computer Aided Modeling 2011. aware vertical handovers,” International Symposium on Applied
[7] J.-M. Chang, W.-T. Hsiao, J.-L. Chen, and H.-C. Chao, “Mobile Sciences in Biomedical and Communication Technologies, ISABEL
Relay Stations Navigation-based Self-Optimization Handover 2009. pp. 1-6.
Mechanism in WiMAX Networks,” IEEE, 2009. [18] R. Yusoff and M. D. Baba, “Performance Analysis of
[8] S.-R. Yang, C.-C. Kao, W.-C. Kan, and T.-C. Shih, “Handoff Transparent and Non-transparent Relays in MMR WiMAX
Minimization Through a Relay Station Grouping Algorithm With Networks,” IEEE Symposium on Industrial Electronics &
Efficient Radio-Resource Scheduling Policies for IEEE 802.16j Applications, ISIEA 2011, Malaysia.
Multihop Relay Networks,” IEEE Transactions on Vehicular [19] A. Laily, Norsuzila, and M. Tarmizi, “Handover Initiation
Technology, vol. 59, 2010. Across Heterogeneous Access Networks for Next Generation
[9] L. Xujie, “A Fast Handover Scheme for Wimax System,” Wireless Cellular Network,” IEEE Symposium on Wireless Technology and
Communications Networking and Mobile Computing WiCOM 2010 Applications ,ISWTA 2010, Malaysia.
© 2012 ACEEE 15
DOI: 01.IJNS.03.02. 39